Quantum Capacity

Sending information in the quantum realm is a delicate balancing act. Unlike classical bits, quantum bits (qubits) are incredibly fragile, their states easily corrupted by interaction with their environment. This inevitable degradation poses a fundamental challenge: how do we quantify the true ability of a noisy physical process, or "quantum channel," to transmit quantum information faithfully? This question highlights a critical gap in simply extending classical information theory to the quantum world, requiring a new and more nuanced measure of a channel's performance. This article introduces this measure: the quantum capacity. Across the following chapters, we will unravel this essential concept. In "Principles and Mechanisms," we will explore the core ideas that define quantum capacity, from the crucial role of coherent information to the conditions under which a channel's ability to transmit quantum data collapses to zero. Subsequently, in "Applications and Interdisciplinary Connections," we will see how this theoretical limit provides a practical blueprint for technologies like the quantum internet and a foundational concept in fields ranging from physics to cryptography.

Imagine trying to send a secret message written in invisible ink. The message itself is delicate, but the real challenge is that the paper it's written on is slightly damp. As the note travels, the ink might not just smudge; it might react with the water, leaving a faint, discolored trace on the blotting paper it was resting on. To understand how much of your secret gets through, you can't just look at the final message; you must also check the blotting paper to see what was lost.

This is the central challenge of sending quantum information. A quantum state, a "qubit," is the invisible ink. The physical process it travels through—a fiber optic cable, a microwave resonator, or even empty space—is the "quantum channel." And just like the damp paper, this channel inevitably interacts with its surroundings, the "environment," causing the pristine quantum state to degrade. The measure of a quantum channel's utility isn't just about how many qubits get through, but about how much of their essential "quantumness" survives the journey. This measure is the ​​quantum capacity​​.

To measure the capacity of a classical channel, like a telephone line, we ask how many distinct bits (0s or 1s) can be reliably distinguished at the output. For a quantum channel, the question is more subtle. We need to quantify the amount of quantum coherence that can be preserved. This is captured by a remarkable quantity called the ​​coherent information​​, denoted $I_c$.

To grasp it, let's personify the concepts. We measure information using ​​entropy​​, $S$, which you can think of as a measure of surprise or uncertainty. If you receive a message that is completely random, its entropy is high—it's full of surprises. If it's a long string of identical letters, its entropy is zero. The coherent information is a simple-looking but profound formula:

$I_c = S(\text{output}) - S(\text{environment})$

This formula tells a story. The potential for sending quantum information, $I_c$, is the "surprise" contained in the state that arrives at the destination, $S(\text{output})$, minus the "surprise" that has leaked into the environment, $S(\text{environment})$. The information you leak is a liability that you must subtract from your gains. You can only successfully transmit quantum information if you can keep your secrets from the environment more effectively than you transmit them to the receiver. The ​​quantum capacity​​, $Q$, of the channel is then the highest possible coherent information you can achieve by cleverly choosing your input state.

Some channels, by their very nature, enforce a stark and unforgiving balance. Consider a hypothetical channel that applies one of two random errors, a Pauli $X$ or $Z$ flip, with equal probability. A deep dive into its mechanics reveals that for any qubit you send through it, the information leaked to the environment is a perfect mirror of the information in the output state. The entropy of the output is always identical to the entropy of the environment, $S(\mathcal{E}(\rho)) = S(\text{environment})$. The balance is always zero, $I_c = 0$. Consequently, its quantum capacity is zero, meaning no quantum information can be reliably sent through it at all.

A channel with zero quantum capacity is not just a poor channel; it's a fundamentally broken one for quantum communication. No matter how ingenious your error-correcting code, you cannot protect a single qubit from its noise. The physical principle that determines this point of no return is as quantum as it gets: entanglement.

Imagine you have a pair of particles in a perfectly entangled state—their fates are intertwined, no matter how far apart they are. You keep one particle and send the other through your quantum channel. If the particle that emerges on the other side is no longer entangled with the one you kept, the channel has "broken" the entanglement. A channel that does this for any entangled state you send through it is called an ​​entanglement-breaking channel​​.

This property is the kiss of death for quantum communication. A fundamental theorem of quantum information theory states that a channel has ​​zero quantum capacity if and only if it is entanglement-breaking​​. Entanglement is the very resource that quantum error correction schemes use to protect information, so a channel that destroys it on contact is hopeless.

The ​​depolarizing channel​​ is a perfect laboratory for this idea. It models a process where, with probability $p$, any incoming qubit is completely scrambled into a state of maximum randomness. For a $d$-dimensional quantum system (a qudit), there exists a critical noise threshold, $p_{th} = \frac{d}{d+1}$. If the noise $p$ is above this value, the channel becomes entanglement-breaking. For a three-level system, or "qutrit" ($d=3$), this threshold is $p_{th} = 3/4$. Any depolarizing qutrit channel with noise greater than $75\%$ has a quantum capacity of exactly zero. This isn't a gradual fade to black; it's a phase transition where the channel's fundamental nature changes, and its ability to convey quantumness ceases to exist. Any channel that can be modeled as a "measure-and-prepare" process—where the input state is effectively measured and a new state is prepared based on the outcome—is also entanglement-breaking, and thus has zero capacity.

To truly understand why a channel fails, we must follow the information that leaks out. Quantum mechanics insists that information is never truly lost, only moved. The description of any quantum evolution, if we zoom out far enough, is a perfectly reversible unitary process acting on both the system and its environment. The channel $\mathcal{E}$ we see is just one side of the coin. The other side is the ​​complementary channel​​, $\tilde{\mathcal{E}}$, which describes the state the environment is left in.

This perspective gives us a powerful new classification:

A channel is ​​degradable​​ if its output is "better" than the environment's. More formally, if we can process the channel's output with another noisy channel to perfectly simulate what the environment saw. This is good news! It means more coherent information stayed in the main channel than leaked out.

A channel is ​​antidegradable​​ if the reverse is true: the environment's version is "better." We could, in principle, process the environment's state to simulate our output. The eavesdropper has the better copy. This is catastrophic for quantum communication. For any antidegradable channel, the coherent information is always less than or equal to zero, meaning its ​​quantum capacity is zero​​.

This brings us to one of the most elegant results in the field, which ties back to the famous ​​no-cloning theorem​​. What is the ultimate eavesdropper? A cloning machine. An optimal symmetric $1 \to 2$ quantum cloner is a physical process that takes one qubit and produces two (imperfect) copies. We can view this as a channel where one clone is the output, and the other is the "leak" to the environment. The properties of this cloner are such that it is perfectly antidegradable. The two clones it produces are of equal fidelity, meaning the environment's information is a perfect match for the receiver's. The delicate balance needed for a positive coherent information is destroyed. The very act of optimal cloning creates a channel with zero quantum capacity.

This principle is remarkably general. Any mixture of zero-capacity antidegradable channels also results in a channel with zero capacity. And the effect is shockingly sensitive. It's even possible for an adversary to take a perfectly good channel—one with positive capacity—and render it useless by applying a tiny, carefully chosen nudge to the state of the environment before the interaction even happens. This perturbation can be enough to tip the channel into the antidegradable regime, killing its capacity entirely. The ability to communicate quantumly is not just a property of the communication line itself, but is inextricably tied to the quietness and stability of its surrounding world.

While many channels are inherently hostile to quantum information, the story doesn't end there. The "capacity" of a channel is not a single, immutable number; it depends on the resources we are allowed to use.

First, let's consider a simple reality of complex systems: noise processes often happen in sequence. If a qubit must pass through channel $\mathcal{N}_1$ and then channel $\mathcal{N}_2$, the total capacity is governed by the ​​data processing inequality​​. It states that the capacity of the composite channel can be no greater than the capacity of its weakest link: $Q(\mathcal{N}_2 \circ \mathcal{N}_1) \le \min\{Q(\mathcal{N}_1), Q(\mathcal{N}_2)\}$. If you have a zero-capacity component anywhere in your system, like an amplitude damping channel with more than $50\%$ energy decay probability, the quantum capacity of the entire chain collapses to zero.

But what if we add resources? Let's look at the ​​erasure channel​​, where a qubit either arrives perfectly or is detectably lost. With a $1/3$ erasure probability, the best we can do with standard error correction gives a capacity of $1/3$. But what if the receiver can call the sender on a classical phone line and say, "Qubit number 5 was lost"? This ​​feedback-assisted communication​​ dramatically changes the game. The sender can simply re-send the lost qubits. The protocol becomes vastly more efficient, and the capacity doubles to $2/3$. The simple addition of a classical feedback loop transforms the channel's potential.

Finally, what about uncertainty? In the real world, we rarely know the exact nature of the noise. Imagine a scenario where the channel is either a bit-flip channel or a phase-flip channel, but we don't know which. We must design a single, robust code that works for both. This ​​compound channel​​ problem seems like it should have a much lower capacity due to the ambiguity. And yet, for this symmetric case, a clever choice of encoding makes the channel's identity irrelevant. The capacity of the compound channel is $1 - H_2(p)$, exactly the same as the capacity of each channel individually. By embracing the uncertainty and designing for it, we lose nothing.

Quantum capacity, then, is a rich and multifaceted concept. It is a story of a delicate balance between a signal and its shadow, a litmus test for the survival of entanglement, and a practical measure of what is possible when we expand our toolkit to include feedback, robustness, and a deeper understanding of the quantum world.

Now that we have grappled with the principles and mechanisms behind quantum capacity, you might be left with a perfectly reasonable question: “So what?” Is this merely a fascinating but esoteric piece of mathematics, a plaything for theorists? Or does it connect to the world we can see, touch, and build? The answer, perhaps satisfyingly, is that quantum capacity is one of those rare ideas that is both profoundly abstract and intensely practical. It is a thread that ties together disparate fields, from the engineering of a future quantum internet to the fundamental principles of cryptography and the microscopic dance of atoms. It acts as a universal Rosetta Stone, allowing us to translate problems from one domain into the language of another, revealing surprising and beautiful unities along the way.

In this chapter, we will embark on a journey to see this principle in action. We won't be deriving more formulas; instead, we'll be exploring the landscape of ideas where quantum capacity is the crucial landmark, guiding our understanding and our aspirations.

First, let us adopt the viewpoint of a physicist. To a physicist, any physical process that transforms a quantum state can be thought of as a "quantum channel." A photon bouncing off a mirror, an atom colliding with its neighbors in a gas, a qubit slowly decohering due to ambient magnetic fields—all of these are channels. The quantum capacity of that process, then, becomes a physical observable. It’s a quantitative measure of how well that process preserves the delicate quantum information of the state. It tells you how “quantum” the interaction is.

Imagine, for instance, a spin-1/2 particle, our qubit, adrift in a sea of other particles. Inevitably, it will suffer collisions, which tend to randomize its spin. This chaotic process is a source of noise. We can, however, model this entire complex, many-body phenomenon as a simple depolarizing channel, where with some probability $p$, the qubit’s state is replaced by a completely random mixed state. The quantum capacity of this channel tells us how much quantum coherence survives this collisional chaos.

The story becomes even more interesting when we realize we can control this process. Suppose we shine a laser on this bath of particles. The laser field can alter the dynamics of the collisions, influencing how they affect the qubit's spin. In the language of channels, changing the laser intensity $I_L$ changes the error probability $p$. This means we can literally tune the quantum capacity of the system with a dial in the lab. We can actively engineer a physical process to be a better or worse preserver of quantum information, turning capacity from a passive property into a controllable feature of the system. This simple idea—that physical processes are channels whose capacities can be measured and controlled—is a powerful tool for analyzing and manipulating the quantum world.

Of course, some channels are so destructive that no quantum information can survive. Consider a channel that, with some probability, simply throws away your input state and replaces it with a fixed, predetermined state, like a freshly prepared entangled pair. While this might sound useful, the fact that the channel sometimes ignores your input completely can be catastrophic for communication. In such cases, the channel’s ability to transmit the original quantum state is fatally compromised, and its quantum capacity can drop to exactly zero. No matter how cleverly you encode your information, you cannot reliably send a quantum message through a channel that occasionally refuses to listen.

While physicists see channels everywhere, engineers want to build them. The most direct and ambitious application of quantum capacity is in the design of the quantum internet—a global network for transmitting quantum information. Quantum capacity is the fundamental "speed limit" that governs every component of this future technology.

Let’s start with the most basic building block: a point-to-point link, like an optical fiber. When we send a quantum state encoded in light down a fiber, two things happen. The signal gets weaker due to absorption and scattering—a process we model as a loss channel with transmissivity $\eta$. To combat this, we might need an amplifier with gain $G$ before sending the signal. But quantum mechanics tells us that amplification is never perfect; it must add its own noise. The composite channel, an amplifier followed by a lossy fiber, is a realistic model for a single link in a quantum network. Its capacity, which depends on the gain $G$, the loss $\eta$, and the amount of energy $N_S$ we are allowed to put into each signal, gives us the ultimate data rate for that link. This calculation is not just academic; it dictates the trade-offs engineers must make when designing repeaters and amplifiers for long-distance quantum communication.

But a true quantum internet will be a complex web of these links, not just a single strand. Imagine a simple "diamond" network, with a source S, a sink T, and two relay nodes in between. Information can travel along two paths. What is the total capacity of this network? It's not simply the capacity of one link. If the relay nodes can perform sophisticated joint quantum operations, they can essentially weave together the information from the two paths. In this case, the network's capacity is governed by a quantum version of the famous max-flow min-cut theorem from classical network theory. The total throughput is limited by the narrowest "cut" through the network, which in this simple symmetric case is the sum of the capacities of the two parallel paths from the source to the relays. The concept of capacity scales from a single link to an entire network, providing the theoretical foundation for routing protocols on the quantum internet.

Real-world engineering also involves dealing with complex, structured noise. A channel's noise is not always a simple, uniform fog.

• ​​Channels with Side-Information:​​ What if a sensor tells us that a particular segment of our fiber network is currently operating under ideal conditions, while another is noisy? This is a "flagged" channel, where we get some classical information about the noise process. In this scenario, we can be clever. We can choose to only use the "good" part of the channel. The total capacity, then, is not the average capacity, but rather the weighted average of the capacities of the available branches. Any extra information about the noise can be exploited to boost our transmission rate.
• ​​Channels with Memory:​​ Noise is not always independent from one moment to the next. A fluctuation in a power supply or a lingering thermal effect can mean that an error at one time step makes an error at the next time step more likely. This is a channel with memory. To find its capacity, we need to understand not just the noise itself, but its dynamics—how the error type evolves over time, perhaps as a random walk. The channel's ultimate capacity is then reduced not only by the average noise but also by the "information" contained in the noise correlations themselves, quantified by the entropy rate of the memory process.

The capacity of a channel tells us the ultimate speed limit. But how does one actually drive at that speed? The answer lies in the art and science of quantum error correction. Channel capacity and error-correcting codes are two sides of the same coin: one describes the problem (the noisy channel), and the other describes the solution (a robust encoding).

The theory of quantum error-correcting codes provides its own set of bounds, like the celebrated quantum Gilbert-Varshamov (GV) bound. This bound guarantees the existence of codes with a certain rate $R$ that can protect against a certain number of errors. It's a statement from a "builder's" perspective: "I can build you a code that is this good." Channel capacity, on the other hand, is a statement from an "information theorist's" perspective: "No matter what anyone builds, you can't go faster than this." It is a beautiful fact that these two perspectives often converge. Under certain reasonable assumptions, the maximum rate promised by the existence theorems for codes perfectly matches the ultimate rate limit given by the channel capacity. The two fields are looking at the same fundamental truth from different angles.

This duality also contains a stern warning, enshrined in what is called the converse of the coding theorem. The capacity is not a mere suggestion; it is a hard wall. What happens if you get greedy and try to transmit information at a rate $R$ that is higher than the channel's capacity $Q$? The converse theorem gives a quantitative answer: you will fail, and your failure will be measurable. For any code operating above capacity, there is a strict lower bound on the error, or equivalently, an upper bound on the fidelity of the transmission. For example, a version of the quantum Fano inequality for a qubit erasure channel shows that the best possible entanglement fidelity $F_{e, \text{max}}$ you can achieve is strictly less than 1, and is precisely limited by the difference between your rate and the capacity: $F_{e, \text{max}} = \frac{R + 1 - 2p}{2R}$. The further you push your rate $R$ above the capacity $Q(p) = 1-2p$, the worse your fidelity gets. There is no free lunch in quantum communication.

Perhaps the most profound and surprising connection is the one between quantum capacity and information security. What does it mean to send a classical message privately? It means that not only does the intended recipient, Bob, get the message, but a malevolent eavesdropper, Eve, who has complete control over the channel's environment and can collect everything that "leaks" out, learns absolutely nothing.

The rate at which you can send such perfectly secure information is called the private capacity, $P$. A stunning result in quantum information theory reveals a deep connection between the ability to send information privately and the ability to send it quantumly. The private capacity $P(\mathcal{N})$ and the quantum capacity $Q(\mathcal{N})$ of a channel $\mathcal{N}$ are fundamentally linked. For a large and important class of "degradable" channels (where the receiver gets a better version of the state than the eavesdropper), these two capacities are exactly equal: $P(\mathcal{N}) = Q(\mathcal{N})$.

This principle allows us to analyze the security of a channel by studying its physical properties. For example, consider channels that are ​​antidegradable​​—where the eavesdropper gets a "better" version of the state than the receiver. Such channels are fundamentally insecure. A physical example arises when the information leaked to the environment is forced into a structure, like the antisymmetric subspace, that makes the main channel itself antidegradable. For any antidegradable channel, both the quantum capacity and the private capacity are identically zero. Not all channels can be used for secure communication! This powerful connection tells us that the key to privacy lies in ensuring the receiver has an information advantage over any potential eavesdropper..

As our journey comes to an end, we see that quantum capacity is far more than a single number. It is a unifying concept, a single mathematical idea that provides:

• A physicist’s measure of coherence in a microscopic system.
• An engineer’s speed limit for a future quantum internet.
• A coder’s benchmark for the performance of error-correcting schemes.
• A cryptographer’s foundation for the very possibility of secure communication.

From the chaos of colliding atoms to the grand blueprint of a global network, quantum capacity offers a common language to describe the flow and preservation of our most fundamental resource: information. Its study reveals the deep and often surprising unity of the physical laws that govern our world, a hallmark of the kind of beautiful, interconnected physics that makes its exploration such an inspiring adventure.